Electron Stagewright docs

ADR-001: Naming and License

Context

Before any code is written, the project needs a name that satisfies four criteria:

  1. Available on npm and GitHub without collision against active projects in the testing / AI-agent automation space.
  2. Trust-signaling to professional developers evaluating tooling for production use.
  3. Domain-anchored — the name must communicate what the project does without explanation.
  4. Future-proof — the name must survive internal architecture pivots (e.g., if the Playwright _electron experimental API is deprecated and the transport layer must be rewritten).

The project also needs an open-source license appropriate for an MCP server targeting both individual developers and enterprise adoption.

Decision: Name

Electron Stagewright — npm package electron-stagewright, GitHub org electron-stagewright.

Rationale

Alternatives considered and rejected

Name Why rejected
Stagehand Direct collision with browserbase/stagehand — 22.8k stars, "The SDK For Browser Agents". Adjacent domain, active development. Catastrophic confusion.
Maestro Direct collision with mobile-dev-inc/maestro — 14.2k stars, "E2E Automation for Mobile and Web". Same domain.
Showrunner Collision with active npm package (Feb 2026) for desktop screen recorder.
Foreman Active npm package, process manager.
Electron Conductor Considered. Strong double meaning (electrical conductor + orchestra conductor) but more generic; loses semantic precision without the prefix.
Electron Sentinel Considered. Strong fit for the production-validation capability set, but biases perception toward a single feature rather than the holistic driver.
Electron Lodestar Considered. Risk of being perceived as pretentious; collision with ChainSafe Lodestar (Ethereum, 1.4k stars) outside our domain.

Availability verified (2026-05-26)

Decision: License

MIT License.

Rationale

Alternatives considered and rejected

License Why rejected
Apache-2.0 Stronger patent grant, used by Playwright and Electron itself. Considered as the primary alternative. Rejected because MIT is the ecosystem default for MCP and the patent grant's marginal value is low for a protocol-implementation project. Easy to revisit if a future legal review surfaces a concrete need.
BSL (Business Source License) Considered for future commercial sustainability (managed cloud session-trace service). Rejected at this stage because (a) the community open-source perception of BSL is mixed, (b) the project has no commercial product to protect, (c) MIT preserves the option to dual-license a future SaaS layer without restricting the current OSS core.
GPL-3.0 / AGPL-3.0 Rejected — would prevent embedding in proprietary downstream tools, which is exactly the adoption path for MCP servers (Claude Code, Cursor, etc. all run third-party MCP servers in proprietary processes).

Consequences

References